AN ANALYSIS OF AIR COMPRESSOR FAULT DIAGNOSIS USING MACHINE LEARNING TECHNIQUE

Machine Fault Diagnosis is an important domain in Mechanical Engineering which concerns about finding fault in the machine parts. Among many techniques to identify and classify the faults, this paper concerns about using machine learning algorithms to distinguish healthy machines fro mtheun healthy machines. In order to distinguish the state of a machine,classification algorithms has to beused. The accuracy of an algorithm depends upon the pattern, that the data set follows. The suitability of the five most commonly used classification algorithm has been discussed. Various transforms can be applied to such sensor data. Here various algorithms have been tested for wave let packet transform. Thea ccuracy of the fit has been measured for all the five algorithms. Hyper-parameter tuning has been done to make the fitbetter.